I was just wondering what are the optimal settings you guys came across. These are mine
Thread concurrency: 32765
Worksize 512
Intensity 19
Vectors: 1
GPU threads: 1
I get about 790 khash/s
Try your worksize as 256 then set your GPU core to 1000mhz and memory to 1500mhz and you should hover around 860kh/s at intensity 19. That is what I've been running for a while.
If you are curious about other settings with that card, check out this link https://litecointalk.org/index.php/topic,6842.90.html